This data set contains data used in the project Measuring Disagreement in Science documented in the following preprint: arXiv:2107.14641. The data set contains the following data: The 65 disagreement queries (signal phrase, filter phrase, valid or not, number of citing sentences returned by the query, SQL query). Of the 65 queries, 23 passed our 80% validity threshold. The 455,625 citing sentences identified by the 23 validated disagreement queries (DOI, sentence sequence number, queries by which the sentence was identified, sentence text). The publications in which the citing sentences occur (DOI, publication year, number of sentences in the main text, main field to which the publication was assigned, meso-level field to which the publication was assigned, inferred gender of the first and last author). The data was retrieved through the Elsevier ScienceDirect API subject to Elsevier’s policy for text and data mining.